Skip to main content

Full text of "Abbeystead explosion : a report of the investigation by the Health and Safety Executive into the explosion on 23 May 1984 at the valve house of the Lune / Wyre Water Transfer Scheme at Abbeystead"

See other formats


View looking towards the north-west corner of Abbeystead Valve House after removal of damaged roof 
beams 



Printed image digitised by the University of Southampton Library Digitisation Unit 



A report by HM Factory Inspectorate 



Health and Safety Executive 



The Abbeystead 
Explosion 



A report of the investigation by the 
Health and Safety Executive into the 
explosion on 23 May 1984 at the Valve 
House of the Lune/Wyre Transfer 
Scheme at Abbeystead 





HER MAJESTY’S STATIONERY OFFICE 



Printed image digitised by the University of Southampton Library Digitisation Unit 



© Crown copyright 1985 
First published 1985 



ISBN 0 11 883795 8 



Printed image digitised by the University of Southampton Library Digitisation Unit 



Contents 



Introduction 1 

Lancashire Conjunctive Use Scheme Design and 
construction 2 

Operations prior to 23 May 1984 8 

Events of 23 May 1984 9 

The investigation 10 
Preliminary findings 10 
A void in the tunnel 10 

The nature and source of the inflammable gas 12 
The method of entry of gas into the Valve House 15 
Source of ignition 15A 

Testing the theories — A simulation of events 18 
The simulation Comments 19 

Conclusions 19 

Recommendations 20 

Appendix List of technical papers 22 



Printed image digitised by the University of Southampton Library Digitisation Unit 



Introduction 



In the evening of Wednesday 23 May 1984, a group of 44 people including 8 employees of the 
North West Water Authority, was assembled in a valve house set into a hillside at the outfall 
end of the Lune/Wyre Transfer Scheme at Abbeystead. 

The visitors, mainly from the local Parish of St Michaels on Wyre, were attending a 
presentation being part of a programme to allay local residents’ anxieties on the effects of the 
installation on the winter flooding of the lower Wyre Valley. 

As part of this presentation, water was to be pumped over the weir regulating the flow of water 
into the Wyre. Shortly after pumping commenced, with the visitors congregated in the valve 
house which was virtually underground, there was an intense flash, followed immediately by an 
explosion causing severe damage to the valve house. Sixteen people were killed; no one escaped 
without injury from the valve house. 

The Health and Safety Commission directed the Health and Safety Executive to investigate 
and to make a special report in accordance with Section 14(2)(a) of the Health and Safety at 
Work etc Act 1974. 



Printed image digitised by the University of Southampton Library Digitisation Unit 



The Lancashire Conjunctive Use 
Scheme 

Design and construction 

1 A proposal to establish a water supply link between 
the Rivers Lune and Wyre was approved in 1971 by the 
water undertakings which later merged to form the 
North West Water Authority (NWWA). It was to be 
part of the Lancashire Conjunctive Use Scheme, which 
was designed to meet the expected water supply require- 
ments of the area, A well known firm of consultants 
specialising in water supply systems, Binnie and 
Partners, advised on the practicability of the scheme 
and was later commissioned to design the scheme. 

2 Design and preparatory work went on from 1971 to 
1975. Because of the outstanding scenic beauty of the 
area it was decided that everything practicable should 
be done to minimise the effect of the scheme on the 
landscape, and to forestall possible objections from 
local residents, landowners, and environmentalists. 
Angling interests led to a decision to abandon of an 
original intention to chlorinate the water entering the 
system. Other design considerations were the security 
of the installation and protection of valves against 
freezing in Winter. Taken together these led to the 
decision that the proposed valve house building at 
Abbeystead should be largely underground. As a result 
of the care taken to preserve the environment, no 
serious objections were raised when a public enquiry 
was held by the Department of the Environment in 
1974. Planning approval was granted in April 1975 
without major conditions, and tenders for construction 
were invited. 

3 Prior to the commencement of tunnelling, enquiries 
made by the designers of the tunnel suggested that 
information available on the geology of the area was 
limited, current Ordnance Survey geological maps 
being based on surveys made in the 1870s. The data 
obtainable was summarised in documents which were 
supplied to prospective tunnellers by the designers of 
the scheme. The tunnel was to be driven through upper 
carboniferous strata of the millstone grit group, with 
lengths at both portals through superficial Pleistocene 
deposits of more regent age. 

4 Consideration was given by the designers to the 
possibility of obtaining further useful information on 
the geology of the area by drilling bore holes along the 
line of the tunnel, but only a few were actually drilled, 
all of them in the vicinity of the ends of the tunnel. 

Their main purpose was to establish the nature of the 
ground at points where there would be little ground 
cover for the tunnel. It was decided that it would not 
be economically viable to drill sufficient bore holes to 
the depth necessary to establish conditions along the 
whole length of the tunnel. The basis for this decision is 
understood to have been the designers’ extensive 
experience of work on other tunnels where numerous 
bore holes were sunk, with results of limited value. An 

2 



independent specialist on site investigations concurred 
with this view, although an official of the Institute of 
Geological Science (now British Geological Survey) 
consulted at the design stage expressed surprise that 
there was to be no bore hole along the tunnel line. His 
concern was primarily that there might be a major in- 
rush of water during construction, rather than that gas 
might be encountered. 

5 No anxiety appears to have been felt at the design 
stage about the possibility of flammable gases being 
encountered during tunnelling. Although the con- 
ditions of contract required the concentration of 
flammable contaminants to be kept below 10% of the 
lower explosive limit and measuring instruments to be 
provided, these conditions are claimed by the designers 
to be routine. 

6 The contract for construction of the Wyresdale 
tunnel and Abbeystead valve house was awarded to 
Edmund Nuttall Limited, who commenced work on 
site at the end of 1975. The designers of the schemes, 
Binnie and Partners, acted as Resident Engineers and 
the North West Water Authority maintained a close 
liaison with the work by appointing one of its technical 
staff as Project Engineer. 

7 No serious problems were encountered during the 
driving of the tunnel. Although traces of flammable 
natural gas were detected, the contractors and consult- 
ing engineers appear to have regarded the tunnel to be 
gas free by normal tunnelling standards. There was a 
considerable inflow of water through the tunnel walls, 
although less than had been feared. This inflow con- 
tinued even after the concrete lining of the tunnel had 
been completed, and at a late stage in the construction 
work the ground water entering the tunnel through 
cracks and porosities in the walls measured at 

13.5 litres/sec (1.16 megalitres/day). An inflow of 
water into deep tunnels is not unusual as the ground 
water surrounding them may be at much more than the 
pressure inside the tunnel. If water flowing through a 
tunnel is at lower pressure than the ground water an in- 
flow through the tunnel walls can be expected even 
when the tunnel is full of water. It is not then a cause 
for concern as the inflow augments the water supply. 

8 Construction work on the Lune/Wyre link was 
completed in the Spring of 1979 and water was first 
pumped through it in June 1979. The responsibilities of 
the consultants and contractors on the site ended on 
the 15 December 1980. 

The layout of the scheme 

9 The Abbeystead Outfall Station, the Wyresdale 
Tunnel and the Lune Pumping Station form the Lune- 
Wyre transfer link of the Lancashire Conjunctive Use 
Scheme. The Scheme is a water supply project aimed at 
meeting an anticipated increase in demand for water in 
the North Western region. 

10 Figure 1 shows in diagrammatic form the layout of 
the Scheme, which allows for the supply of water from 



Printed image digitised by the University of Southampton Library Digitisation Unit 




Figure 1. General layout of Lune/Wyre water transfer scheme 



3 



3A 



Printed image digitised by the University of Southampton Library Digitisation Unit 









existing impounding reservoirs, by extraction from the 
River Wyre at Garstang and from boreholes sunk into 
the Fylde Bunter sandstone. On occasions .during the 
year there is insufficient water in the River Wyre to 
satisfy abstraction requirements and on these occasions 
it is intended that if sufficient water is available in the 
River Lune it should be transferred to the River Wyre 
and subsequently abstracted downstream at Franklaw. 
It is the Lune-Wyre transfer portion of the Lancashire 
Conjunctive Use Scheme with which this report is 
concerned and the Wyresdale Tunnel, and Abbeystead 
Valve House in particular. (See Fig 1.) 

1 1 The Lune-Wyre transfer link comprises the Lune 
Intake and Screenhouse, the Lune Pumping Station, 
the Quernmore Pipeline, the Wyresdale Tunnel and the 
Abbeystead Outfall Station. 

12 The water intake for the transfer link is situated 
approximately 150 metres upstream of Forge Weir near 
the village of Halton on the River Lune. A short pipe 
leads from the intake bays to the Screenhouse where 
any coarse debris is screened out and from there the 
water flows by gravity along a tunnel 750 metres long 
to the Pumping Station. 

13 The Lune Pumping Station is sited in Quernmore 
Park. Six low lift pumps lift the water up to ground 
level, where it is retained in settling tanks to allow any 
sediment present to settle out. From the settling tanks 
the water flows over weirs into a balancing tank prior 
to transfer to the River Wyre. A further six pumps 
installed on the floor of the pumphouse can deliver a 
nominal 545 Ml/day of water into the Quernmore 
Pipeline, but the maximum amount the NWWA is 
permitted to abstract from the Lune is 280 Ml/ day. 

14 The Quernmore Pipeline runs for a distance of 
5 kilometres from the Lune Pumping Station to the 
entrance to the Wyresdale Tunnel at Rowton Portal. 
The pipeline is approximately 1800mm in internal 
diameter, and is of welded steel construction. It is 
buried at a shallow depth and follows the contours of 
the landscape. Automatic air valves at all high points 
allow air to escape when the line is being filled and 
allow air to enter when the line is being emptied. Wash- 
outs (drain valves) at each low point allow the pipeline 
to be emptied of water. 

15 From Rowton the water is conveyed from the end 
of the pipeline to Abbeystead Outfall Station through 
Wyresdale Tunnel, which is 6.6 kilometres in length. 
The tunnel is concrete lined and has a finished internal 
diameter of 2600mm. At each end of the tunnel there is 
a length of a few hundred metres with a steel lining, 
this being to strengthen the tunnel where the depth of 
the cover is relatively small. This strengthening lining is 
not necessary under current operating conditions but is 
designed to allow operations at greater water pressure 
in a later stage in the development of the scheme. The 
tunnel rises a vertical distance of 5.5 metres up a very 
gentle gradient between Rowton and Abbeystead, the 
highest point in the line being at the end of the access 
branch at Abbeystead. (See Fig 2.) 

4 



The Abbeystead Outfall Station 
17 Abbeystead Outfall Station is situated about 47m 
from the river bank in a rural area of the Wyre Valley 
about 10 kilometres South East of Lancaster. When the 
Wyresdale Tunnel reaches Abbeystead it is continued 
by two branches of 2600mm diameter concrete-lined 
steel pipeline, as illustrated in Fig 2. The access branch 
extends to the tunnel entrance chamber where it 
terminates in a blank flange. Near the end of the access 
branch the tunnel is vented into a buried vent chamber. 
As there are no intermediate access points along the 
length of the tunnel from Rowton to Abbeystead this is 
the main vent point/air intake for the whole length of 
the tunnel. Within the vent chamber eight air valves 
fitted above the pipeline allow air to escape while the 
line is being filled with water and to enter while the line 
is being emptied. The air valves are located within a 
concrete-lined chamber 4.4m X 3.6m X 2.15m high 
which is set into the ground and roofed with earth- 
covered concrete beams to form an underground 
chamber with no ventilation to the open air. An 
800mm diameter steel vent pipe which is permanently 
open leads directly from the vent chamber to the valve 
house, so effectively venting the tunnel into the valve 
house. This vent pipe runs almost horizontally into the 
valve house where it curves downwards to discharge 
just above the grid floor level. There is also a small 
drainpipe from the vent chamber discharging into the 
valve house. As can be seen in Fig 2, the discharge 
branch diverges from the side of the access branch and 
runs at a downward gradient under the lower part of 
the valve house and on beneath the River Wyre, 
terminating in a blank end under the far bank. The dis- 
charge branch continues beneath the river solely to 
allow for any future extension of the scheme. A 200mm 
diameter washout pipe runs from the discharge branch 
to an outfall chamber on the near bank of the river. 

This pipe is fitted with a valve which can be opened to 
allow drainage of the discharge branch of the tunnel, 
but is intended to be kept closed in normal operation. 
There are also 100mm diameter circulation pipes 
extending back from the dead ends of both branches of 
the tunnel and discharging into one of the distribution 
chambers in the valve house. 

17 From the crown of the discharge branch below the 
dry room of the valve house the water is conveyed 
through two 1000mm diameter steel pipe bends (known 
as ‘lobster backs’) and discharged into two discharge 
chambers in the wet room of the valve house (see Figs 
3a and 3b). The lobster backs discharge water below 
the water level in the chambers and are submerged 
when the chambers are full. A third lobster back, con- 
nected to the discharge branch, consists of 800mm 
diameter steel pipework discharging into a nearby 
hollow in the surrounding field, and is intended as an 
emergency discharge line. The three lobster backs are 
fitted with butterfly valves with controls located in the 
dry room. For normal running the control valve on the 
emergency discharge lobster back is kept closed and the 
control valves on the other two lobster backs are kept 



Printed image digitised by the University of Southampton Library Digitisation Unit 



Abbeystead Portal 



7- 






m m 



\ 



\ 






■'"v\ 

\ 



x\ 



% XV 
\\ 

\ 



a, 



a®* 






Discharge branch 77 metres 
from Abbeystead Portal 



s\ \\ 






\\ 

« - 



\\ ' 



w 



E ”«se^ d -- : 



l5ch ^h,-;~— ^ — }} 

\ \\ — 

\Vfev 1 



Tunnel vent 
chamber 




'%'k 



Figure 2. Abbeystead Outfall Station 



Printed image digitised by the University of Southampton Library Digitisation Unit 



open. There is an air valve similar to those in the vent 
chamber on each of the lobster backs venting air just 
below floor level in the wet room. 

18 When a discharge chamber is full, the water flows 
over a weir, each discharge chamber supplying 2 of 4 
distribution chambers. A little water can also enter the 
distribution chambers through the 100mm diameter cir- 
culation pipes leading back from each of the dead ends. 
Four outfall pipes convey the water from the distri- 
bution chambers to outfall chambers on both banks of 
the river. Outfall pipes 1 and 3 feed the outfall 
chambers on the near bank and pipes 2 and 4 run 



beneath the River Wyre and feed the far bank outfall 
chambers. The final discharge into the river should 
therefore be reasonably uniformly distributed between 
both banks in normal operation. The discharge ports 
from these outfall chambers are permanently open. The 
downstream chamber on the near bank is supplied with 
water from the washout valve referred to in paragraph 
16, and when the valve is opened water from this 
chamber enters the river through the five discharge 
ports at the downstream end of the outfall. 

19 The construction of the Valve House at Abbey- 
stead was of reinforced concrete and it was predomi- 




Access branch 



Fig 3a Illustration of valve house, tunnel vent chamber and tunnel access chamber 



6 



Printed image digitised by the University of Southampton Library Digitisation Unit 



Ventilation panels 





Plan of valve house with roof and floor omitted for clarity 



Fig 3b Valve House 

7 



Printed image digitised by the University of Southampton Library Digitisation Unit 



nantly underground, only the wall at the entrance side 

being above ground. The roof consisted partly of solid 

concrete and partly of precast concrete beams 

400 X 450mm in cross-section, and was overlaid with a 

waterproofing treatment and a layer of top soil and 

turf. 

20 Entry to the Valve House from outside was 
through substantial outward opening double doors on 
to an intermediate floor above the tunnel and the dis- 
charge chambers. These doors led into a room 
12.5m long X 5.15m wide and 2.5m high known as the 
dry room, which was directly above the discharge 
branch of the tunnel. The dry room housed the 
controls for the water valves on the lobster backs, 
instruments and a toilet compartment. It had a floor 
consisting partly of solid concrete with an inset section 
of steel mesh over the valve pits. A second pair of sub- 
stantial double doors in an internal dividing wall 
opened from the dry room into a room known as the 
wet room. This room was 9.50m long X 7.50m wide by 
2.50m high and occupied the whole area above the dis- 
charge and distribution chambers. Its floors consisted 
entirely of steel mesh panels carried on joists 8.30m 
above the bottoms of the discharge chambers. 

21 Ventilation for the wet room was provided by a 
single louvred ventilation panel measuring 1410 by 
880mm set in the outside wall. There was no mechan- 
ical ventilation. In the dry room, air was admitted 
through a 600 X 330mm louvred opening set near the 
top of the outside wall. A small electric fan extracted 
air through metal trunking from below the floor at the 
rear of the room, where the instruments were situated, 
and discharged through a separate louvred opening 
identical to the inlet. In the dry room there were also 
two small tubular electric heaters which were normally 
kept switched on. 

22 The designers of the tunnel have expressed the 
view, which has been confirmed by enquiries to Water 
Authorities, that no other installation in Great Britain 
incorporates all the characteristics of the Wyresdale 
tunnel and Abbeystead Valve House. The specific 
characteristics which appear to have been of import- 
ance in relation to the Abbeystead explosion were: 

(a) All the contents of the tunnel, both liquid and 
gaseous, discharged into a room with limited 
natural ventilation, 

(b) Water passed through a concrete lined tunnel, i.e. a 
tunnel not designed to be watertight. 

(c) Ground water from the strata surrounding the 
tunnel leaked in rather than tunnel water leaking 
out. 

Operations prior to 23 May 1984 

23 Water was first transferred through the Lune- 
Wyre Link and discharged at Abbeystead on 25 June 
1979. Since then the system has remained operationally 
available, except when it was drained for inspection 

8 



and maintenance in September 1979. Operational con- 
trol was at first vested in the Ribble Division of the 
North West Water Authority, but following a re- 
organisation in 1983 control passed to the Ribble Dis- 
trict of the new Northern Division Supply and Treat- 
ment Department. 

24 Demand for water has proved lower than was fore- 
cast when the Scheme was planned, so less use has been 
made of the Lune-Wyre Link than was expected. 
Records show that between 1 January 1983 and the 
date of the explosion, pumping took place on 246 of 
the 509 available days, and there was no pumping for 
several weeks on a number of occasions. The maximum 
quantify of water pumped in any one day was only 

87 megalitres, as against the 280 megalitres/day which 
the Authority is permitted to extract from the Lune. 

25 It was a feature of the design that during the 
standstill periods the tunnel would remain full of water, 
so long as all washout valves through which it could be 
drained were kept closed and no leaks developed. 
Indeed, providing there was no drop in the inflow of 
ground water which was known to enter through the 
tunnel walls, there should normally have been a con- 
tinuous flow of water at all times over the weirs in the 
Valve House and out into the river through all the dis- 
charge ports, even if no pumping of water was taking 
place. 

26 The Manual of Operating Instructions provided by 
the designers of the Scheme indicates that the washout 
valves at Abbeystead and Rowton Portal should be 
opened periodically to flush out ‘dead water’ from the 
system. No interval between openings is specified but 
the practice of flushing out monthly was adopted 
originally by the NWWA. Normally this was done 
while pumping was in progress, when opening of the 
washouts would have no noticeable effect on levels of 
water in the system. Any significant opening of a 
washout valve while pumping was not in progress 
might have resulted in partial drainage of the tunnel, 
but when the valve was re-closed, the system should 
soon have been re-filled by the inflow of ground water, 
and flow through the valve house should have restarted. 

27 A period of 25 days continuous pumping was com- 
pleted on 6 May 1984, this being the last time the 
system was used before the explosion occurred. 

Towards the end of that period of pumping, routine 
visits were paid to the Abbeystead site by the Group 
Superintendent who was operationally responsible for 
it, and by his assistant, both of whom normally visited 
about once a month. Throughout the pumping period 
and the subsequent 17 days before the explosion, 
routine calls were also made by an operator who nor- 
mally visited the site for about 1 to 2 hours weekly to 
keep the plant and installations clean and tidy. The 
same operator was responsible for the monthly atten- 
tion to the washout valve referred to above, but it was 
not a specified part of his duties to check the water 
flow in the valve house during his weekly visits. 



Printed image digitised by the University of Southampton Library Digitisation Unit 



28 The last routine visits to the site prior to the 
explosion were paid on the day on which it occurred, 

23 May 1984, The operator paid his weekly visit 
around 1.00 pm, and at about 2.15 pm a technician 
from the Authority’s River Division called to check the 
calibration of a meter installed in the dry side of the 
valve house. Neither of them went into the wet room 
where water is discharged, to check the flow of water at 
the weirs. 

29 Except when the routine visits referred to in the 
previous paragraph were being paid, the Abbeystead 
installation was normally unattended at all times. 

There was no operational need for any person to be 
present when pumping was started or stopped. A 
number of parties of visitors had been shown round the 
valve house, since the commissioning of the Scheme, 
some while pumping was in progress, but it has not 
been possible to establish whether or not pumping had 
ever commenced while visitors were present. 

30 During the construction of the Wyresdale tunnel 
and the Abbeystead outfall, HM Factory Inspectorate 
inspected the civil engineering and construction work 
on a number of occasions. After the scheme had 
become operational the Factory Inspectorate made a 
general inspection of the Abbeystead outfall instal- 
lations in July 1981. Records of these visits make no 
reference to the presence of flammable gases. 

The events of 23 May 1984 

3 1 Between 6.30 pm and 7.00 pm on the fine, dry 
evening of 23 May 1984, a party of 44 visitors 
assembled at the Abbeystead Valve House. Thirty-six 
of the visitors were residents of the village of 

St Michael’s-on-Wyre, about 15 miles away, where 
there had been concern that floods in the village during 
recent years might have been aggravated by water 
pumped into the Wyre at Abbeystead. The other 8 
members of the party were employees of the North 
West Water Authority. Six of them were members of 
the technical staff concerned with the operation of the 
Lancashire Conjunctive Use Scheme, who were to act 
as guides, the other 2 being office workers who were to 
assist with refreshments. 

32 When the visit began no water was being pumped, 
and none had been pumped for 17 days, but it was in- 
tended that pumping should take place during the visit 
for demonstration purposes. After giving an introduc- 
tory talk outside the entrance to the Valve House, the 
North West Water Authority’s District Supply and 
Treatment Manager gave orders over a telephone in- 
stalled in the dry room for a pump to be started at the 
Lune Pumping Station at the supply end of the Lune- 
Wyre Link. A pump which had a nominal capacity of 
37.5 Ml/day was switched on at about 7.12 pm and 
pumping of water from the balancing tank at Quern- 
more into the tunnel began. Meanwhile, the visiting 
party had split up. Some people went into the Valve 



House, others down to the nearby riverbank where the 
water outfall from Abbeystead entered the Wyre 
through the rows of discharge ports on both banks of 
the river. Visitors who went to the riverbank saw that 
water was flowing from only a few of the discharge 
ports, at the downstream end of the outfall on the near 
bank of the river, i.e. those fed by the washout valve in 
the discharging branch. Those who went into the wet 
room of the valve house noticed that there was no flow 
of water into the discharge chambers, and that the 
water levels in the chambers were well below the weirs. 

33 When no water had flowed into the Valve House 
after about 10 minutes a second telephone call was 
made to the pumping station by the District Manager 
and an order was given for a second, variable speed 
pump to be started, to double the rate of flow to 

80 Ml/ day. By then almost all the visiting party had 
gone inside the Valve House. 

34 Just before 7.30 pm, when water had still not 
started to flow into the Valve House, an explosion 
occurred inside the House. Every member of the 
visiting party was either killed or injured, many of 
them suffering severe burns. Eight people died at the 
scene and the death toll rose eventually to 16, including 
3 of the North West Water Authority Officials who 
were acting as guides. Substantial damage was caused 
to the Valve House. Part of the earth-covered concrete 
beam section of the roof was lifted by the explosion 
and collapsed into the valve house, causing some of the 
more serious injuries. The collapsing beams broke 
through the steel mesh flooring of the wet room and 
some of the injured fell into the water in the chambers 
below. 

35 The explosion was heard by a number of people in 
the surrounding countryside, some of whom reached 
the scene of the accident within 10 minutes and 
attempted to help survivors, at some risk to themselves. 
The emergency services were alerted within 5 minutes, 
but their stations were several miles away, and it was 
7.48 pm when the first ambulance arrived on site, 
followed almost immediately by the other emergency 
services. 

36 The first flow of water into the valve house was 
noticed by some of the survivors several minutes after 
the explosion, probably between 7.40 and 7.50 pm. At 
that time the pumps were still running, and they con- 
tinued to run until the pump operator at Lune Pump- 
ing Station telephoned Abbeystead to check, on the 
water flow at about 7.50 pm. A local resident answered 
the telephone and told the operator that there had been 
an explosion. Suspecting a hoax the operator tele- 
phoned to the main control room at Franklaw Water 
Treatment Works for confirmation and it was not until 
8.05 pm that the pumps were switched off. Rescue 
operations continued throughout the night until about 
6.00 am, when all known visitors had been accounted 
for. HM Inspectors of Factories (HMFI) from North 
West Area and the North West Field Consultant 



Printed image digitised by the University of Southampton Library Digitisation Unit 



Group were on site by 1.30 am on the 24 May and 
established contact with the emergency services and the 
management of NWWA, before commencing their 
investigation. 

37 A first inspection of the interior of the Valve 
House was made early on 24 May 1984. At that stage 
much of the roof had already beep removed by the 
rescue services but it was possible to see that several of 
the large concrete beams forming part of the roof had 
collapsed inwards and that most of the beamed section 
of the roof had been lifted from its supports. The 
damage was greatest in the wet room where a number 
of roof beams had fallen onto the steel mesh floor and 
broken through sections of it into the discharge and 
distribution chambers below. One of the roof beams in 
the dry room had also collapsed onto the floor between 
the entrance to the Valve House and the doorway to 
the wet room. The internal doors and their surrounds 
had been blown out of position, as had the louvred 
ventilation panel for the wet room which was found 
later on the opposite bank of the river. Signs of smoke 
and flame were very limited, except in a small area just 
inside and to the left of the main entrance, where there 
were clear signs of burning near a partially buried 
cupboard. Elsewhere the walls appeared quite clean 
except for adherent fragments of what seemed to be 
burnt clothing and slight spatterings of bitumen-like 
material in places. The general effect was consistent 
with an explosion of a gas-air mixture having occurred 
inside the Valve House. There were also clear signs of a 
similar explosion in the nearby vent chamber, its earth- 
covered roof having sustained severe damage. 

The investigation 

Preliminary findings 

38 In the initial stages of the investigation attention 
was directed towards a number of potential expla- 
nations for the explosion, in addition to the presence of 
naturally occurring methane. These included the 
possible storage of liquefied petroleum gases or 
flammable liquids in the Valve House, entry into the 
Valve House of commercial gas or other flammable 
substances leaking from pipelines in the area, the 
electrolytic generation of hydrogen in the tunnel and 
terrorist activity. It soom became apparent that most of 
these potential explanations could be eliminated but 
one or two were thought to merit some further 
investigation. 

39 The position of all relevant pipelines in the area 
was established and with the assistance of the British 
Gas Corporation all gas lines were checked and tested, 
no evidence being found of any leakage which might 
have reached Abbeystead Valve House. Neither was 
there any evidence of the storage of gas cylinders on 
the premises, but a 5 gal drum buried by fallen debris 
in the area where fire damage had been noticed just 
inside the Valve House entrance proved to have con- 

10 



tained a very small quantity of petroleum spirit. This 
drum was situated close to a small wall-mounted 
tubular heater, which could have raised its temperature, 
and there was some evidence that the contents had 
ignited and burst the drum. The drum was examined by 
experts from the North West Field Consultant Group 
of HMFI and the Health and Safety Executive’s 
Research Laboratory Services Division (RLSD), who 
concluded that the petroleum spirit was likely to have 
been vaporised and ignited by the heat generated by the 
explosion. Accordingly the drum was discounted as a 
possible source of the explosion. The strong probability 
which emerged was that the explosion had been caused 
by the ignition of gas, likely to be mainly methane, 
which had accumulated in a void in the tunnel, and 
that the pumping of water had displaced flammable gas 
from the tunnel into the Valve House. 

40 Although other possibilities were kept in mind, 
subsequent investigation centred on the following 
questions. 

(a) How, when, and to what extent had a void arisen 
in the tunnel? 

(b) What were the nature and source(s) of any gases 
present in that void? 

(c) How had the gas been transferred to the Valve 
House? 

(d) How and where had the gas been ignited? 

All these lines of enquiry are considered separately 
below. 

A void in the tunnel 

41 The Lune-Wyre Link was designed to run full of 
water during pumping and to remain full with the 
pumps switched off. In these circumstances no signifi- 
cant void could exist in the tunnel during normal 
operation. If the integrity of the system was maintained 
and none of the washout valves along the tunnel line 
was open, water should never have been below weir 
level in the Valve House, and providing the inflow of 
ground water into the tunnel was at its normal level, 
there should have been a small continuous outflow of 
water over the weirs in the Valve House and through 
all the 52 normal discharge ports on both banks of the 
river. 

42 The evidence from victims of the explosion indi- 
cated that no water was flowing into the discharge 
chambers in the Valve House and that the water level 
in them was approximately 1 metre below the weirs 
prior to the explosion. The only water flowing into the 
river was from the five downstream discharge ports on 
the near bank of the river, i.e. the ports supplied only 
from the washout valve at Abbeystead. 

43 An outflow of water from the five downstream 
outfall ports alone could occur only if the washout 
valve on the near bank of the river was at least partially 
opened. The low water level and absence of flow at the 
weirs indicated that the water level in the system had 
fallen sufficiently for water to have ceased to flow 



Printed image digitised by the University of Southampton Library Digitisation Unit 



through the lobster back discharge pipes. This 
suggested that the washout valve was sufficiently open 
to drain off more than the inflow of ground water into 
the tunnel. This would have resulted in the creation of 

<s> 

a void, the size and speed of development of which 
would depend on the difference between the rate of 
inflow into the tunnel and of outflow through the 
washout, assuming that there was no substantial 
leakage elsewhere in the system. 

44 There appeared to be a number of other factors 
which might have contributed to a void in the tunnel, 
for example a decrease in the water inflow, faulty wash- 
out or air valves along the line, incorrect opening of 
those washout valves or failure to close them, abstrac- 
tion of water during the prevailing dry weather by un- 
authorised opening of washout valves, or leakage from 
the system as a result of structural faults or damage. A 
detailed internal examination of the tunnel was clearly 
essential. It had already been drained back to just 
beyond Rowton Portal when a washout valve there was 
opened in the course of the rescue operations, but 
before it could be entered great caution had to be 
exercised because of the risk that flammable concen- 
trations of gas might still exist or arise. 

45 Assistance was sought from HM Inspectors of 
Mines and Quarries who have knowledge and experi- 
ence of work in methane-rich atmospheres, and of the 
Mines Rescue Service of the National Coal Board. 
Scientific staff from the HSE’s RLSD which incorpor- 
ates the former Safety in Mines Research Establish- 
ment, also joined the investigating team at this stage, 
and continued to play an active part throughout the 
investigation. 

46 When the tunnel was first entered, the presence of 
methane was soon detected at levels rising to 2% of the 
atmosphere at 1.5 km from the access point at Abbey- 
stead. Although the methane concentrations were 
below the lower flammable limit of 5%, they were 
above the level at which work would be permitted in 
coal mining practice, so before the investigation inside 
the tunnel could continue it was necessary to ventilate 
the tunnel thoroughly. The large blank flange at the 
end of the access branch of the tunnel at Abbeystead 
was therefore removed and with the assistance of the 
National Coal Board, who made available a large 
extraction fan and the services of an experienced mine 
ventilation engineer, a satisfactory extraction venti- 
lating system was installed within a few days and 
progress could be made with the internal examination 
of the tunnel. 

47 A visual examination of the inside of the tunnel 
revealed no major defects in the lining. Inflows of 
ground water through fissures and porous areas were 
noted at various points in the concrete lined centre 
section, but measurements of the total inflow using a 
temporary weir at Rowton indicated that it was of the 
same order as the flow when the tunnel was taken into 
use, i.e. approx 1 megalitre/day (11 to 12 litres/sec). 



48 Various ‘tide marks’ were seen on the walls of the 
tunnel at the Abbeystead end. It was at first thought 
that they might indicate water levels at different times 
and in different operating conditions. A number of 
theories were advanced to account for some of these 
tide lines but no explanation has been found for them 
which contributes to the picture of events in the tunnel 
leading up to the explosion. Another curious 
observation was that stalactite-like pendant growths 
had formed at certain points which would be expected 
to be permanently submerged. It was at first suggested 
that these growths could not form under water and so a 
large void must have existed for long periods. This 
theory was rejected after research by the Common- 
wealth Mycological Institute established that the 
growths could form under water. 

49 ‘Tide marks’ were also found in the washout 
chamber at the river outfall and these were thought to 
indicate water levels at differing settings of the washout 
valve. The results of tests based on water levels in the 
washout chamber carried out during the investigation 
are reported in greater detail in para 94. 

50 While water sampling and investigations into the 
pattern of water flow in the tunnel were being carried 
out by Inspectors from North West Field Consultant 
Group, parallel enquiries were going on into the oper- 
ating procedures followed by the NWWA, particular 
attention being paid to use of the washout valve at 
Abbeystead. The interviewing of NWWA staff directly 
concerned with the operation of that valve revealed 
that the practice of opening it periodically, as 
recommended in the Operating Instrutions, had been 
abandoned early in 1980. This change followed 
excessive silting up of the dead-end under the river at 
Abbeystead which led to severe discolouration of the 
river when the valve was opened. 

51 Early in 1980 it is understood to have become the 
practice to open the washout valve, probably V/ 2 to 

2 turns of the spindle whenever pumping was in 
progress, and to allow the discharge of water through 
the washout valve to continue until pumping ceased, at 
which stage the washout valve was closed. After a few 
months, when pumping through the Lune-Wyre came 
under remote control from Franklaw Treatment 
Works, the method of using the washout was changed 
again. It is said to have been left permanently ‘cracked 
open’, from then onwards, whether pumping was in 
progress or not. The setting described as ‘cracked open’ 
appears to have been about three-quarters to one turn 
of the washout valve from the fully shut position. It has 
not proved possible to establish who authorised the 
changes in the procedure for using the washout valve at 
Abbeystead. No documentation relating to the changes 
has been produced, and no guidance on their potential 
effects appears to have Teen sought from the designers 
of the system. 

52 It seems probable that the changes were intro- 
duced by NWWA operating staff who saw no reason 




Printed image digitised by the University of Southampton Library Digitisation Unit 



for referring them to a higher level in the organisation 
for approval. If this had been done the possibility of 
partial drainage of the tunnel might have been recog- 
nised, and it may well be that the procedures would 
have been modified to prevent any potential loss of 
water. No evidence has come to light to suggest that 
the presence of flammable gas in the tunnel had been 
envisaged by anyone concerned with the operation of 
the scheme. 

53 With the washout valve open approximately one 
turn, water was found by the plant operators at the 
time of change of procedure to continue to flow over 
the weirs in the Valve House and out through all the 
outflow ports, including the washout ports. 

54 Initially, following the change in procedure, it was 
the practice for the operator to check the flow over the 
weirs when he paid his routine weekly visit to Abbey- 
stead Outfall, but with the passage of time and an 
eventual change of operators, the regular checking of 
flow over the weirs was abandoned and only the flow 
of water from the outfall ports at the river was 
checked. The operator who carried out the checks from 
November 1983 until the explosion has stated that he 
saw water flowing from the end five outfall ports sup- 
plied by the washout valve whenever he checked, but 
he is uncertain about the flow of water from the 
remaining ports. The fact that an absence of flow from 
these remaining ports indicated that water was not 
flowing over the weir was not appreciated by him. 
None of the more senior members of the surviving 
NWWA staff who visited the site occasionally noticed 
water flowing from only the end five outfall ports. 

55 Information on the flow of water through the out- 
fall ports into the river was also obtained from two 
independent witnesses not in the employment of the 
NWWA, one of them an angler who fished regularly 
over the Abbeystead stretch of the Wyre and the other 
a local gamekeeper. Both these witnesses have stated 
that there were occasions when no water was flowing 
into the river through the outfall ports. In particular, 
the gamekeeper reports visiting the outfall very 
regularly for a few weeks before the explosion and 
seeing no water runing. The angler, who was fishing the 
river when the explosion occurred, reports walking 
down the riverbank on the opposite side to the Valve 
House about 1% hours before the explosion and 
noticing water coming out of about four jets at the 
downstream end on the Valve House bank of the river, 
i.e. the outfall ports supplied by the washout valve. He 
also noticed a similar situation on the weekend before 
the explosion. About two weeks before the explosion 
he saw no water flowing into the river on two nights. 

56 The evidence of the angler referred to in the 
previous paragraph confirmed that of several survivors 
of the explosion. It appeared to establish beyond 
reasonable doubt that immediately before the explosion 
occurred, and probably for some days before that, 
water had been flowing into the River Wyre only 

12 



through the outfall ports supplied by the washout 
valve. This would indicate the absence of any flow over 
the weirs in the Valve House that was reported by 
several survivors of the explosion. The absence of any 
flow into the river reported by both the witnesses 
referred to in the previous paras is harder to explain 
but is considered below in para 59. 

57 The fact that water was flowing only from the 
washout ports did not necessarily establish that there 
was a substantial void in the tunnel, but this was 
clearly indicated by the lapse of time between the 
starting of pumping on 23 May and the arrival of water 
in the Valve House some time after the explosion. If 
the void had resulted solely from a loss of water at the 
Abbeystead washout the maximum void would have 
arisen when the water in the tunnel was only just flow- 
ing into the discharge branch. The water level would 
then have reached the tunnel crown approximately 
291m from the Abbeystead portal and a void of 
approximately 1425m 3 (1.425M1) would have existed. 

58 The size of any void could have been increased if 
any of the other washout valves along the line of the 
pipeline between Quernmore and Rowton had been 
opened, either officially or by persons seeking to 
extract water for their own purposes during the dry 
spell preceding the explosion. 

59 The opening of these valves could also have led to 
the absence of flow at Abbeystead described by the 
witnesses referred to in para 56 above. Enquiries into 
this possibility proved fruitless, however, and no 
information has come to light suggesting that any such 
unauthorised abstraction occurred during the standstill 
period before the explosion. 

60 The probability appeared to be that the primary 
cause for the creation of a void was the departure from 
the instruction in the operating Manual represented by 
the leaving of the Abbeystead washout valve partly 
open. Providing that valve was set in the position 
described as normal by the NWWA employees con- 
cerned, it would not in itself have drained off sufficient 
water to create a void. There is a possibility that a leak 
somewhere in the system may have contributed to the 
drainage (and could also have led to a complete 
stoppage of the water flow at Abbeystead) but no sig- 
nificant leak has been identified and it seems unlikely 
that any exists. It follows that the washout valve was 
set in a more open position than had been described as 
normal, or that there had been unauthorised inter- 
ference with the operation of the system. No evidence 
of interference has been found. 

Nature and source of the inflammable gas 

61 The first positive evidence of the presence of 
methane was obtained from the analysis of water 
samples taken when the tunnel was drained down at 
the Rowton end soon after the explosion. These 
samples were found to contain concentrations of dis- 
solved methane of from 5 to 9 mg/litre. Samples of 



Printed image digitised by the University of Southampton Library Digitisation Unit 



water taken from the pipework between Rowton and 
Quernmore were found to contain less than 
0.5 mg/litre, so it was inferred that any source of 
methane was unlikely to be in the Lune to Rowton part 
of the system, or in water extracted from the Lune. 

62 Methane was detected in the atmosphere when the 
Abbeystead end of the tunnel was first entered two 
days after the explosion and tests were carried out with 
a methanometer. With no ventilating plant installed, 
the concentration of methane in air reached 2% about 
1.5 km from Abbeystead, 5.2 km from Rowton Portal 
and it appeared that most of the methane was likely to 
be originating somewhat nearer the Rowton end than 
that point. 

63 Methane is a flammable gas which, if confined, 
can give rise to an explosion. A combustible mixture of 
flammable gas and air can exist only between certain 
limits known as the lower and upper flammable 
(explosive) limits. A mixture below the flammable limit 
is too weak to burn and a mixture above the upper 
flammable limit is too rich to burn. The flammability 
limits for methane are 5% and 15% by volume in air. 

64 Although traces of other gases were detected they 
were not present in significant quantities in the tunnel 
atmosphere, so subsequent investigation was centred 
on identifying the source of methane, and the way in 
which it had entered or been generated in the tunnel. 
Initially there appeared to be two likely hypotheses 
which will be referred to subsequently as the biological 
and geological theories: the methane may either have 
been generated in the tunnel by anaerobic decompo- 
sition of organic matter in the water, or it may have 
entered the tunnel from fissures in the surrounding 
strata through cracks or porous sections of the tunnel 
wall. There was also the possibility that a mixture of 
methane from both sources was present. 

65 Considering first evidence for the biological 
theory; there were deposits of blackish watery sludge 
several inches deep along the bottom of the tunnel, and 
in the dead end dipping below the river what appeared 
to be a large deposit of sludge had built up at a point 
just beyond the inlets to the lobster backed pipes 
supplying the Valve House. When the dead end of the 
tunnel was subsequently pumped out, the water content 
of this sludge was found to be high and the total mass 
of solid matter relatively low. The sludge was regarded 
as one potential source of methane. Another was a wet 
slime which had been deposited on the tunnel walls. 
Advice on their methane-generating potential was 
sought from the Rowett Research Institute of 
Aberdeen, whose sampling and subsequent culture 
growth revealed the presence of methanogenic 
organisms. Their research also indicated, however, that 
the quantity of methane likely to have been generated 
could not be expected to be sufficient to create the 
explosive atmosphere which had existed. 

66 To check the validity of the geological theory of 
the origin of the methane it was necessary to establish 



the age of gas samples from the tunnel. Samples were 
taken from the atmosphere at various points and sub- 
mitted to the Isotope Measurements Laboratory at 
Harwell for radio-isotope dating. The results indicated 
that the methane was ancient (i.e. in excess of 20000 
years old) and so must be presumed to be mainly of 
geological origin. The possibility of a mixture of 
methane of different ages could not be excluded, but 
the result of the Rowett Institute tests described in the 
previous paragraph suggested that the content of recent 
biological methane was low. 

67 Water entering the tunnel by leaking through the 
walls, and water lying in and flowing through the 
system was also tested and analysed. At normal tem- 
perature and pressure, methane is not very soluble in 
water, but if the pressure in the surrounding strata was 
high its solubility would be raised significantly. 

Samples of water entering the tunnel taken at various 
points showed methane contents ranging from a trace 
up to 40 mg/litre. As this water entered the tunnel and 
was exposed to a lower ambient pressure some of the 
methane would be given off immediately as gas, par- 
ticularly where the solution was super saturated and if 
discharge was into a length of tunnel not full of water. 

If the tunnel was full of water, methane bubbles would 
be likely to form in the crown of the tunnel and would 
tend to move up the slope, much of it dissolving in the 
water, and eventually enter any void existing in the 
Abbeystead end of the tunnel. Thus the methane con- 
tent of the tunnel water would be raised by the 
methane-rich inflow, as was suggested by the samples 
of tunnel water taken during the draining of the tunnel 
and at later stages in the investigation. 

68 The visible flow of water into the tunnel was 
entirely in the concrete lined section. The main inflow 
occurred well beyond the end of the steel, lined sections 
of the tunnel at a distance of between approximately 
4.2 and 5.7 km from Rowton Portal (1.0 and 2.5 km 
from the Abbeystead end). In that area several small 
jets of water could be seen emerging from the tunnel 
wall and there were also clear signs of more diffuse 
leakage. Because of difficulties in sampling and 
measuring the amount of ingress water it was not 
possible to estimate how much methane was entering 
the tunnel dissolved in water. 

69 Measurements of low concentrations of methane 
in air extracted from the empty tunnel by the venti- 
lation system installed during the investigation indi- 
cated that the total quantity of methane entering or 
being given off in the tunnel at that time was 
O.OOlmVsec. This was the only measurement of the 
total quantity of methane in gaseous form that it was 
possible to carry out. It must be regarded as only 
approximate, because of the limitations in measuring 
both low concentrations of gas and the volumetric air 
flow. 

70 It was concluded, following the completion of all 
the tests described and a detailed inspection of the 

13 



Printed image digitised by the University of Southampton Library Digitisation Unit 



tunnel, that the flammable gas which was present was 
almost entirely methane and was mainly of geological 
origin. Most of it was leaking through the tunnel walls 
in solution in water between 4.0 and 5.5 km from 
Rowton Portal, where it had already been established 
that the tunnel was likely to have been full of water 
prior to the explosion. 

71 Efforts were made to reconcile these findings on 
the geological origin of the methane with information 
which had come to light about conditions prior to and 
during construction of the tunnel. Present knowledge 
of the geology of the area, based partly on surveys 
carried out by the British Geological Survey during the 
driving of the tunnel, is shown diagrammatically in Fig 
3c. It will be noted that the tunnel may have passed 
through the top of a sandstone anticline, a structure 
resembling an inverted dome, at about 4 to 4.5 km 
from Rowton (2.2 to 2.7 km from Abbeystead), this 
being in the area where the greatest levels of methane 
in ingress water have been measured. Truncated limbs 
of other anticlinal structures, in the form of bands of 
sandstone, also cross the tunnel in other areas where 
water is entering the tunnel in quantity. The sandstone 
anticline and bands are likely to act as aquifers through 
which water can flow in the strata, possibly carrying 
methane from a distant source in solution. 

72 Local inhabitants have reported that the presence 
of methane near the surface of the ground in areas near 
Abbeystead has been known for many years. This 
appears to have been biogenically produced methane of 
the marsh gas or ‘will-of-the-wisp’ variety rather than 
the methane of ancient origin detected in the tunnel 
during the investigation. Methane has also been 
detected in the spoil heaps at Abbeystead and Rowton 
made up of materials excavated during construction of 
the scheme but the significance of its presence is 
difficult to assess. The spoil is a mixture of material 
excavated at various points, some perhaps near normal 
ground level. Some of it is known to have been contami- 
nated during the construction period by organic sub- 
stances, including diesel oil which might now be giving 
off methane. Finally, results of a borehole which indi- 
cated some presence of hydrocarbons, sunk by Place 
Oil & Gas in 1966 some 3 km from the line of the 
tunnel (the Whitmore borehole) had been lodged with 
the British Geological Survey under ‘commercial in 
confidence’ cover, but were not available. 

73 During driving of the tunnel a good deal of 
ground water was encountered although less than had 
been feared. Evidence of the presence of flammable gas 
during tunnelling is far less clear. Although Draeger 
gas detectors were provided, as required by the 
contract, they were evidently used to a very limited 
extent. They can in any case give only a qualitative 
indication of the presence of natural gas because of the 
sensitivity of the indicator tubes to other gases known 
to have been present at the construction stage. Tests 
are known to have been done on 10 occasions, all of 
them between May and October 1977. The tunnel 

14 



ventilating plant was in operation during all but one of 
these tests. The results of all these tests have been 
traced and examined. Three showed indications of the 
possible presence of natural gas in small quantities, so 
advice on their interpretation was sought from the 
manufacturers of the measuring instrument. HSE’s 
investigating team concluded that the results of the 
tests could not be taken to give a reliable indication of 
either the presence or absence of methane in the 
atmosphere of the tunnel at the time the tests were 
carried out. 

74 No incidents involving naturally occurring 
flammable gas are known to have occurred during con- 
struction, although the tunnel was driven by drilling 
and explosive blasting, the electrical plant used was 
not flameproof, welding and cutting was done during 
the installation of the steel linings at each end of the 
tunnel, and a fire involving a transformer occurred 
inside the tunnel near the Rowton Portal without any 
indication of the ignition of unexpected gases. At one 
stage of the construction, a dead end extended for 
about 1 km beyond the ventilation trunking at the 
Rowton end of the tunnel, reaching as far as the section 
where the highest methane levels have been measured 
in ingress water during the investigation. No incidents 
were reported to have occurred in this dead end, 
although some of the potential means of ignition 
referred to above are said to have been present in the 
unventilated length. Largely on the strength of this 
experience, members of the staff of the designers and of 
NWWA present during the construction, believe that it 
is extremely improbable that methane could have been 
entering the tunnel then, to even a fraction of the 
present extent. But the tunnel was normally well venti- 
lated and in the absence of comprehensive testing for 
flammable gas, it is impossible to reach a firm con- 
clusion on the basis of their observations and any 
inferences drawn from them. 

75 A former employee of the contractor, who worked 
as a ganger in the tunnel, has reported that on three 
occasions he suffered headaches which from previous 
mining experience he associated with the presence of 
methane. However, it has proved impossible to cor- 
relate his recollection of the places where his headaches 
occurred with surviving records, or with the existing 
pattern of water and gas ingress in the tunnel. Head- 
aches often result from other types of fumes which are 
encountered during tunnelling or from oxygen 
deficiency in poorly ventilated conditions. 

76 Experts on relevant aspects of geology have been 
consulted and various theories have been put forward 
to explain the presence and nature of geological 
methane in the strata and its mode of migration into 
the tunnel. On the evidence available, it has not proved 
possible to reach any firm conclusion. Advances in 
geological knowledge which have arisen from the wide- 
spread search for oil and gas in this country in recent 
years, have introduced one new factor, which would 
not have been recognised when the tunnel was designed, 



Printed image digitised by the University of Southampton Library Digitisation Unit 



<r 



*. 



♦ 





R 1 — R a Reference letters of sandstone beds 
on the Rowton limb of the anticline 

A 1 — A 5 Reference letters of sandstone beds 

on the Abbeystead limb of the anticline. 



but which is emphasised by the discovery of a natural 
gas field in the nearby Morecambe Bay. It is now con- 
sidered possible that an underground reservoir of 
natural gas might have existed in the vicinity of the 
tunnel, and that this reservoir may have been tapped in 
some way as a result of changes in the ground water 
channels in the strata initiated by the driving of the 
tunnel. There is also a possibility, suggested by major 
differences in the chemical analyses of water inflows at 
different points, that more than one source of methane- 
containing water exists and that the sources function in 
different ways. The explanations for the existence of 
methane which have received consideration are in 
simple terms. 

(a) Continuous emission of methane of the firedamp 
type familiar to miners and its transfer by ground 
water in Wyresdale Fell. 

(b) The existence of a low pressure methane reservoir 
in Wyresdale Fell. 

(c) The existence of a high pressure methane reservoir 
in Wyresdale Fell. 

(d) A flow of methane, mainly in solution, from a deep 
high pressure source possibly at some distance 
from the tunnel. 

(e) In association perhaps with one of the above, the 
existence of a pocket of methane which was largely 
discharged on the night of the explosion. 

77 Because of the uncertainty which now exists, both 
as to the presence of methane at the time of the design 
and construction of the tunnel, its source and the likely 
pattern of its production at the time of the explosion, 
now and in the future, the North West Water Authority 
is sponsoring further geological investigation in the 
area surrounding the tunnel. 

The method of entry of gas into the Valve House 

78 It was established at an early stage in the investi- 
gation that a substantial void existed in the tunnel on 
23 May 1984. The atmosphere in that void could have 
been transferred to the Valve House in several ways. 
Providing the standing water level in the discharge 



Fig 3c Horizontal section along the line of the Wyresdale Tunnel. 
Except for the 1000m south of the Rowton Portal, the extrapolation 
of the sandstone units above and below the tunnel is conjectural, and 
has been made to illustrate structural relationships. Broken lines are 
used to denote uncertainty. 

chambers was above the level of the outlets of the 
lobster backs in the chamber, as indicated by the 
evidence of the survivors, the primary routes for the 
flow of air and gas would be from the tunnel, through 
the eight air valves in the vent chamber near the Valve 
House, and then via the 800mm diameter vent pipe 
from the vent chamber into the wet room of the Valve 
House. The secondary route would be through the air 
valves on the lobster backs and directly into the wet 
room. Some air/gas might also pass into the wet room 
through the drain pipe from the vent chamber which 
discharges into one of the distribution chambers, and 
also through the circulation pipe at the access dead-end 
of the tunnel beyond the Valve House, but the 
quantities would be relatively small. The main flow of 
gas would be through the air valves in the vent chamber. 

79 The air flow could be in either direction; from the 
tunnel into the Valve House, or vice versa. While the 
tunnel was draining, air from the outside atmosphere 
would be drawn into it through the air valves in the 
vent chamber and the Valve House, which fresh air 
could enter through the ventilation louvres in the 
external wall. While the water was standing, the move- 
ment of air might be in either direction, depending on 
atmospheric conditions, but it would probably not be 
of great significance. Once pumping restarted air/gas 
would be displaced from the tunnel as the level of 
water rose in it and would pass through the air valves 
in the vent chamber and on the lobster backs. 

80 When the water level rose sufficiently to seal the 
lobster backs, the flow of air/gas through the valves on 
them would cease and the flow would continue only 
through the valves in the vent chamber. The faster the 
rate of pumping the greater would be the rate of flow 
of air/gas, which would fill the vent chamber and pass 
through the large vent pipe into the valve house, 

15 



81 There appeared, therefore, to be an easily identifi- 
able gaseous flow mechanism by which any flammable 
atmosphere inside the tunnel could be transferred to 
the Valve House, although it was difficult to predict 
accurately the exact distribution of flow, between the 
different routes. 

Source of ignition 

82 From the evidence of survivors it is clear that the 
explosion originated in the wet room of the Valve 
House, when between 15 and 20 people were in that 
room. Their individual recollections vary in detail, but 
there is a common thread of a flash probably around 
the floor level near the north-west corner of the wet 
room, followed by a blast of burning gas through the 
doorway into the dry room and out through the main 
entrance doors of the Valve House. Six of the eight 
bodies recovered from the scene of the explosion were 
found in the wet room and there was a preponderance 
of structural damage in that same area. 

83 The probable mode of ignition of the gas cloud 
has received careful and detailed consideration but a 
precise source has not been established although a 
number of possibilities have been identified. In a room 
where electrical equipment was not intrinsically safe, 
flameproof or pressurised, and where over 40 people, 
some known to be smokers, were present, ignition by 
one means or another was highly probable if an explos- 
ive atmosphere arose. 

84 The electrical equipment in the wet room where 
the explosion originated, consisted solely of weather- 
proof light fittings and their associated circuitry, all 
instruments, heaters, switches and other electrical gear 
being in the outer dry room. The whole installation was 
subjected to thorough examination by one of HM 



Electrical Inspectors of Factories and specimens of the 
light fittings and other items of electrical equipment 
were later examined and tested in the HSE’s labora- 
tories. Their reports indicate that there were no faults 
in the electrical equipment that were likely to have 
provided a source of ignition. A possibility exists of 
sparking caused by current tracking between damp 
terminals but there were no positive signs of it having 
occurred. The measured operating temperatures of the 
lamps, including the two large floodlights installed 
beneath the grid floor in the wet room and of the 
tubular heaters in the dry room were well below the 
temperature at which ignition of methane could occur. 

85 Items of smoking equipment have been recovered 
from the Valve House, and although they might have 
been there before the day of the explosion, it is possible 
that some of them were the property of the visitors on 
23 May. A lighted cigarette or pipe might not ignite 
methane but the use of a cigarette lighter or match 
would. One survivor reports seeing a man lighting a 
pipe some minutes before the explosion, and it would 
not be surprising if several people lit cigarettes or pipes 
during the rather lengthy wait for water to come 
through the system. No blame could rest on anyone 
who did, for visitors were not warned against smoking 
in the Valve House, the NWWA not having envisaged 
that a flammable atmosphere could arise there. 

86 Another possible source of ignition which cannot 
be excluded is the spark discharge of static electric 
charges built up on clothing. A report by HM Electrical 
Inspector indicates that the removal of a garment made 
of synthetic fibre by a person inside the wet room 
might possibly have generated sufficient energy to 
initiate the explosion. 



15A 



Printed image digitised by the University of Southampton Library Digitisation Unit 



On 



Valve Tunnel 




Figure 4 Abbeystead 

Schematic arrangement of outlet works showing water levels 



Printed image digitised by the University of Southampton Library Digitisation Unit 



Valve Tunnel 

house vent 




Figure S Abbeystead 

Schematic arrangement of outlet works showing gas flow 



Printed image digitised by the University of Southampton Library Digitisation Unit 



Testing the theories — a simulation of 
events 

87 As a result of the investigations a likely scenario 
seemed to have been established for the explosion, so it 
was decided to put it to the test by staging a simulation 
of the events prior to the explosion. The system would 
be re-filled with water and left standing for the same 
period as it stood before 23 May, and then pumping at 
the same rate as on the night of the explosion would be 
re-started under test conditions. 

88 After all those concerned with the various aspects 
of the investigation had been given a final opportunity 
to carry out any further tests considered necessary 
inside the tunnel, preparation for the simulation began 
early in July. 

89 Temporary roofs were erected for the Valve House 
and vent chamber and new ventilation louvres and 
internal doors were fitted to replace the ones severely 
damaged by the explosion. Inside the tunnel, and in the 
Valve House and vent chamber, sampling points for 
gas and water were installed and connected to remotely 
situated measuring instruments. The tunnel ventilation 
fan installed during the investigation was disconnected 
and the blank flange and the manhole cover at the 
access end of the tunnel at Abbeystead were replaced. 
After all personnel had been withdrawn to a safe 
distance and all foreseeable sources of ignition 
eliminated, pumping was recommenced in mid July for 
the first time since the explosion. 

90 The equivalent of two fillings of the system was 
pumped through it to flush out the water which had 
been stagnant in the pipeline end of the Lune/Wyre 
link since the time of the explosion and to reproduce so 
far as was practicable the conditions when pumping 
ceased on 6 May 1984. 

91 After re-filling, pumping was stopped with all the 
washout valves in the system closed and locked off. 
Water continued to flow over the weirs in the valve 
house as a result of ground water entering the tunnel. 
This inflow had been measured by checking the flow 
over a weir at Rowton Portal while the tunnel was 
empty. It was found to be 0.9 to 1.1 Ml/day, a rate of 
flow similar to that measured when the tunnel was com- 
missioned. A subsequent measurement of water flowing 
from the washout ports at Abbeystead indicated that 
with the tunnel full the inflow was only 0.69 Ml/day, a 
reduction of about 30% and a more recent measure- 
ment following wet weather was 0.79 Ml/day. 

92 On the morning of 23 July, the washout valve at 
Abbeystead was re-opened one full turn, which was the 
extent to which it had been the practice to open it prior 
to the explosion, according to the NWWA employees 
most directly concerned. Within a few minutes the flow 
of water over the weirs ceased. The only flow into the 
river was from the 5 end ports at the downstream end 

18 



of the outfall on the near bank, the ports supplied 
through the washout valve. This was the condition 
observed by witnesses immediately prior to the 
explosion. 

93 The flow of water over the weirs did not recom- 
mence, but within a short time the level of water in the 
discharge tanks stopped falling and settled only inches 
below weir level. A gauge installed at the access end of 
the tunnel indicated that the level in the system had 
settled at just above the level of the tunnel crown, at 
that point which meant that the system was still effec- 
tively full of water. In these circumstances no void 
would have developed in the tunnel, the outflow of 
water through the washout being balanced by the 
inflow of ground water through the tunnel walls. 

94 The level of water in the chamber fed by the 
Abbeystead washout valve was compared with a 
number of ‘tide-marks’ noticed on the walls of the 
chamber. The marks were at 23mm, 35mm, 53mm and 
75mm above the bottom level of the outfall ports 
through which water flowed into the river, and were 
thought to indicate levels of flow. In particular the 
HSE investigating team considers that the 53mm mark 
represents the setting of the washout valve used for 
several months before the explosion. At this setting the 
outflow through the washout is believed to have just 
balanced the inflow of ground water when the valve 
was set. With the washout valve setting of 23 July, i.e. 
open one complete turn, the water settled at 44mm 
above the bottom level. As this was within the range of 
tide marks it was considered appropriate to maintain 
the valve setting and observe developments. 

95 By 27 July there had been no significant change in 
water levels, and it was evident that no significant void 
was going to develop in the system at the existing valve 
setting in the absence of any unpredictable events. As it 
was clear that a void had developed between May 6 
and 23, it was decided to increase the outflow through 
the washout sufficiently to ensure partial drainage of 
the system. 

96 Accordingly the Abbeystead washout valve was 
opened a further 5 / g of a turn on 27 July, making a total 
of l 5 / 8 turns from the closed position. The level in the 
washout chamber rose to 70mm above the outfall port 
level, i.e. 5mm below the highest ‘tide-mark’ in the 
chamber. At this setting the water level in the tunnel 
started to fall and by 8 August a void of considerable 
size had formed. However, it was thought possible that 
it was smaller than the void on 23 May, since the water 
level had been falling for only 12 days since the re- 
setting of the washout valve, as against a possible 

17 days before the explosion. 

97 Between 23 July and 8 August methane levels in 
the system were monitored regularly. The results 
showed a methane in air concentration of about 9% in 
the small space between the water level and the crown 
of the tunnel near the inner end of the void in the 
access branch at Abbeystead. A representative sample 



Printed image digitised by the University of Southampton Library Digitisation Unit 



from this area taken on 7 August was found to contain 
methane 8%, ethane 0.05%, Carbon dioxide 0.14%. 
Carbon monoxide 0.0001%, Oxygen 17.80% and 
Hydrogen 0.01%. Lower methane concentrations of 3 
to 4% were recorded in the tunnel nearer the Valve 
House and in the vent chamber. In the Valve House 
itself they were generally below 0.5%. 

98 On 8 August pumping of water was recommenced, 
an attempt being made to follow the same sequence of 
operation as on the night of the explosion, with the 
same operator controlling the pumps 

99 It did not prove possible to duplicate exactly the 
pumping pattern of 23 May but the overall pumping 
rate was similar. The exact size of the void which 
existed on 8 August was the equivalent of 1.38M1 as 
indicated by the amount of water pumped by the time 
of the first flow of water over the weir. It has not been 
possible to establish exactly when that occurred on 23 
May but there is strong evidence that it was between 27 
and 37 minutes after pumping started. On 8 August the 
flow started after 35 minutes, so it is possible that the 
void w'hich existed in the tunnel on 23 May was slightly 
less than 1.38M1 measured on 8 August, rather than 
greater as had been thought possible in view of the 
shorter drainage period referred to in para 86 above. 
The order of size of the void on 23 May was clear, 
however. 

100 While pumping was in progress the measuring of 
methane concentrations and water levels in the system 
continued. After the pumps had run for the same 
length of time as they ran on 23 May before the ex- 
plosion occurred, i.e. approximately 17 minutes, the 
concentration of methane in the atmosphere of the 
Valve House was approximately 3%. This concen- 
tration is non-flammable, the flammable range for 
methane being approximately 5 to 15%. The concen- 
tration in the vent chamber at the same time was 
approximately 4%, also below the flammable range. 

101 A maximum concentration of 7% methane was 
recorded near the roof of the vent chamber after about 
35 minutes pumping. The maximum concentration 
measured in the Valve House at the same time was 
4.80% methane in air, i.e. still just below the flammable 
limit. It was however, fairly uniform at all the sampling 
points in the Valve House, all of which were at or 
above the level of the floor grids. 

The simulation Comments 

102 The simulation cannot have been an exact one. It 
was necessary to go beyond what was stated to be the 
normal setting of the washout valve to create a void, 
and when pumping was resumed an explosive atmos- 
phere was not created in the Valve House. It did 
establish that a substantial void could be created by 
opening the Abbeystead washout valve only slighly 
more than the reported setting, and that a methane-air 
mixture would accumulate in that void. Pumping of 
water would transfer that mixture to the Valve House 



mainly via the vent chamber and an approximately 
uniform mixture of air and methane would form above 
the Valve House floor. The pattern was similar to that 
thought to have led to the explosion but the concen- 
tration of methane was less than had been produced 
between 6 and 23 May. Possible reasons for this were 
thought to be differences in the degree of ventilation of 
the tunnel system, short-term changes in the level of the 
water table in the surrounding ground as a result of pro- 
longed dry weather during the investigation, or a fall in 
the ‘supply’ pressure arising from unknown changes in 
conditions in the strata. It has also been suggested that 
a pocket of gas in the ground may have discharged com- 
pletely into the tunnel on the night of the explosion. 

Conclusions 

103 The explosion was caused by ignition of a 
mixture of methane and air which had accumulated in 
the wet room of Abbeystead Valve House. The 
methane had been displaced from a void which had 
formed in the Abbeystead end of the Wyresdale Tunnel 
during a period of 17 days before the explosion when 
no water was pumped through the system. When pump- 
ing was resumed on 23 May 1984, while the visiting 
party was inside the Valve House, the water level in the 
tunnel rose slowly and pushed out the methane and air 
mixture in the void through the air valves near the end 
of the tunnel. Most of the air/gas mixture passed 
through the eight air valves mounted above the access 
end of the tunnel into a closed vent chamber, from 
where it flowed through a large open vent pipe into the 
Valve House. 

104 The design of the Lune-Wyre Link was novel in 
certain respects for the water supply industry. Its most 
unusual feature in relation to the explosion was the dis- 
charge of water and vented air from the tunnel into an 
enclosed Valve House situated below ground. Although 
ventilation grids were installed in the external wall of 
the Valve House, they were not so positioned nor of 
such a size as to disperse the volume of gas which arose 
on the night of the explosion. If a water discharge sys- 
tem open to atmosphere had been used, an explosion 
would almost certainly not have occurred. 

105 No source of ignition for the explosion has been 
positively identified. Thorough examination and testing 
of the electrical equipment has not revealed any faults 
likely to have caused ignition and there is insufficient 
evidence to confirm any of the other explanations 
which have been considered. Smoking in the Valve 
House was not prohibited because the likelihood of a 
flammable atmosphere arising there had not been 
envisaged. 

106 The void in the tunnel was produced by loss of 
water through a washout valve at Abbeystead which 
had been left permanently open to minimise silt 
accumulation in the end of the tunnels beyond the 
Valve House. A maximum loss of water, approximately 

19 



Printed image digitised by the University of Southampton Library Digitisation Unit 



1.5 Ml, was possible through this washout although the 
evidence indicates that the actual loss at the time of the 
explosion was between 0.97 Ml and 1.42 Ml. A void of 
the same order of magnitude would be created in the 
tunnel, which had been designed to remain full of water 
at all times, except when drained intentionally for main- 
tenance or inspection. 

107 Use of the washout valve in the way described in 
the previous paragraph was not in accordance with the 
operating manual provided by the designers of the sys- 
tem. It had been introduced a year or two after the 
system became operational, apparently without the 
knowledge or approvel of senior operational and tech- 
nical staff employed by the NWWA. If more detailed 
operating instructions had been issued and stricter 
working procedures followed the change in the use of 
the washout valve might have been detected, but the 
possible disastrous consequences of the new procedure 
would almost certainly not have been recognised by 
anyone operationally involved. 

108 Almost all the methane was of ancient geological 
origin, although small quantities probably arose from 
decomposition of organic matter in the tunnel. Most of 
the methane percolated in through the concrete walls of 
the tunnel between 2 and 2.5km from Abbeystead, 
either in a gaseous form or in solution in water under 
pressure. The fact that significant quantities of methane 
might be dissolved in water does not appear to have 
been recognised by the personnel concerned with the 
design and operation of the Lancashire Conjunctive 
Use Scheme, and probably not by the water industry 
generally. Information obtained in the course of the 
enquiry indicates that it has been widely regarded as an 
insoluble gas, its solubility under normal atmospheric 
conditions being very low. References to the presence 
of dissolved methane in water supply systems have 
been traced in published literature but they do not 
appear to have achieved wide circulation, particularly 
amongst the sections of the Civil Engineering 
Profession concerned with water supply schemes. 

109 On the strength of their experience during the 
driving of the tunnel, both Binnie and Partners and the 
NWWA believed, and still believe, that methane was 
not emerging from the strata in quantities which were 
significant. During the construction period, some test- 
ing for methane was carried out in accordance with 
standard contract clauses for safety in the construction 
operation. The results of those tests which were carried 
out showed on three occasions what might have been 
very low levels of methane. But the instrument used 
was susceptible to other gases likely to be present and 
the readings therefore cannot be regarded as confirm- 
ing the presence of methane; in addition the majority of 
the tests were carried out when forced draught venti- 
lation was in operation. 



Recommendations 

1 10 The recommendations concerning design, con- 
struction and operation are intended to apply to 
TUNNELLED RAW WATER TRANSFER SYSTEMS 
which are not of watertight construction. They also 
apply to closed raw water transfer systems in which 
methane is liable to be generated. They do not apply to 
treated water distribution systems using watertight 
pipes. 

Design and construction 

1 1 1 Systems conveying water should be so designed 
that any air or gas discharged, either during filling or at 
any other time, is vented to a safe place in the open air. 

1 12 Where it proves impracticable to comply with 
Recommendation 1 in a particular case, comprehensive 
tests should be made to ascertain the nature of any 
contaminants which might enter the system or be 
generated in it, and appropriate precautions should be 
incorporated in the design to deal with them. 

113 The controls for washout valves should either be 
so located that they cannot be operated by unauthor- 
ised persons or should incorporate arrangements for 
the valves to be locked in the closed position. 

1 14 During tunnelling work, sufficient tests should be 
carried out at frequent and regular intervals to 
establish the presence or otherwise of flammable gases 
using instruments able to provide a quantitative 
reading of acceptable accuracy. On completion, and 
before the commissioning of any tunnel forming part of 
a water transfer scheme, further tests for flammable gas 
should be carried out with no ventilating system in 
operation. The results of such tests should be recorded 
and retained; positive results should be reported imme- 
diately to the designers of the installation and to the 
organisation which will be responsible for its operation. 

Operation 

1 15 Operators of existing raw water transfer systems 
should review the possibility of methane being present 
in the system and, where appropriate, should consult 
with the designers in assessing the safety of the instal- 
lation, particularly where significant voids may be 
formed in the system. 

1 16 Safe systems of work covering all aspects of oper- 
ation and maintenance should be laid down in compre- 
hensive operational instructions, the observance of 
which should be monitored by management. 

117 The training of operating, technical and super- 
visory staff concerned with water transfer systems 
should include measures to ensure that they are made 
fully aware of the significance of any special features of 
the installation with which they are concerned, any 
potential hazards which may be anticipated in the 
course of operations and the appropriate precautions 
to deal with the hazard. 



20 



Printed image digitised by the University of Southampton Library Digitisation Unit 



General 

118 The fact that methane is soluble in water and 
increasingly so above ambient pressure, and that it can 
be given off by ground water entering workings, should 
be widely publicised throughout the civil engineering 
profession and incorporated in professional training 
courses. 

Action by HSE 

1 19 Following the Abbeystead explosion, HSE wrote 



to water authorities alerting them to the possible 
dangers of water transfer and comparable systems 
where methane may create a risk. 

120 Interim advice on the risk of gas evolution at 
water boreholes has been produced for the guidance of 
users. 

121 Further advice on the wider aspects of methane 
evolution from ground water is being prepared. 



21 



Printed image digitised by the University of Southampton Library Digitisation Unit 



Appendix 

The Health and Safety Executive is indebted to the 
following organisations for producing papers which 
assisted with the investigation, and in the preparation 
of this report. 

British Geological Survey: Fluid Processes Research 

Group 

Keyworth 

Nottingham 

Isotopic and Chemical Analyses of Gas and Associated 
Water. Wyresdale Tunnel. 

Commonwealth Mycological Institute 

Culture Collection & Industrial Services 

Ferry Lane 

Kew 

Surrey 

Investigation of structures of possible biological origin 
in concrete pipe at Abbeystead. 



Rowett Research Institute 

Bucksbarn 

Aberdeen 

Two reports on samples taken from the Abbeystead 
Tunnel. 

Harwell: Low Level Measurements Laboratory 

Measurement of Carbon-14 activity in two samples of 
Ba C0 3 supplied as extracted from CH 4 gas collected 
at the Abbeystead Pumping Station. 

In addition to site investigation work the Research and 
Laboratory Services Division of the HSE produced the 
following reports: 

Observations on the damage and deductions as to the 
course of the explosion. 

Examination of and tests on electrical equipment. 

Identification of the fuel, location of its ingress and 
simulation of its displacement from the tunnel. 



Printed in the UK for HMSO 
Dd 715226 C60 2/85 



22 



Printed image digitised by the University of Southampton Library Digitisation Unit